The present invention relates in general to information processing, and more specifically to a system which facilitates dynamic allocation and de-allocation of computing resources such as digital processors, networks, storage devices and software to create a virtual computing environment using an arrangement of networked processors and processes.
Today, computers are increasingly being used in almost every area of commerce, education, entertainment and productivity. With the growing popularity of the Internet, corporate and campus intranets, home networking and other networks, the trend is to use multiple computers, or processing platforms, to perform tasks and provide services. The networked processors approach is in contrast to traditional approaches of running a single application as a stand-alone application on a single computer.
One approach to networked processing is a “client-server” model whereby a server computer on the Internet is used to transfer information to a client computer. Typically, the client computer is located at an end user's location, such as a personal computer in a user's home. This allows large amounts of information to be stored in, and accessed from, the server computer by many client computers. The client computers can access the server computer simultaneously. Another approach allows a user to obtain portions of executable programs from the server to operate an application program in functional “pieces” or components, on the client computer. For example, a user can run a word-processing program in a client-server mode where the server provides portions of the word-processing software to the user's computer on an as-needed basis.
Yet another model is the Application Service Provider (ASP) model. This removes the application still further from the end-user and can involve more processors, such as servers. The ASP model allows a primary server to host a client-server application, or to host any type of data-processing resource such as a database, user interface, program component, data object, etc. The application can appear to the client as being hosted by the primary server when it is actually being provided by one or more other servers. The other servers can provide the application, or components, by having the client directly access the other server, or having the client access the other server through the primary server.
Still other models are possible. Any arbitrary arrangement of processors can be used to provide a service or function to an end user, or to another device such as another processor, consumer electronic device, etc. Examples of such systems are groups of computers used to perform a large mathematical task, such as rendering frames of a video sequence or performing a simulation. Systems that use many processors to exchange information, such as email systems, multicasting programs, voice-over-IP communications, etc., can potentially require hundreds, or even thousands or more, networked processors.
Software is necessary to operate each of these models. Each of the systems described above will likely use multiple software processes on each processing platform. Most likely, the different software processes are designed by different software manufacturers. Although there are standard specifications to design compatible software, such software is still designed largely independently of other manufacturers' software. For example, a user may be running an operating system and web browser designed by different companies. Additional processes may be concurrently executing on the user's computer.
Thus, many independently-designed hardware devices and software processes must be coordinated in order for the overall networked system to operate correctly and effectively. This complexity is multiplied by the number of users that the system is designed to support. For example, where the system is an Internet system it is common for the user base to be in the hundreds of thousands, or millions. There may be hundreds or thousands (or more) simultaneous users operating client processing systems that require tens, or even hundreds, of servers for adequate bandwidth. Different servers in the networked system may perform different functions such as web page serving, database maintenance, caching, etc.
This complexity further manifests itself in another aspect. In the past, software applications used to be monolithic. That is, the application was self-contained within a box with no interaction with other applications, where a box is a computing device or computing machine. However, this paradigm is no longer true. With Internet and e-commerce applications, for example, a client application is configured to communicate with a remotely located server application, such as eBay.com®, or Amazon.com®, or the like. Typically, e-commerce applications define a three-tier architecture that includes: (1) client; (2) server; and (3) database. Within this architecture, the client may be running an Intel® box with a Microsoft Windows X86® environment, the server may be running a Sun Microsystems® box running an iPlanet™ application, the database may be Sun Microsystems) box running Oracle®, for example. Therefore, this three-tiered architecture increases the complexity and cost of many software applications. In one aspect, the complexity of today's computing environments increases computing costs, and extends the time for a new product to reach the market.
Major problems arise where companies need to design, develop, and test software and hardware for use in such large and complex systems. Because today's applications are multi-tiered, it is no longer possible to develop or to test an application within a box. Not only must compatible hardware be selected, obtained, connected and tested; but each piece of software must be selected in accordance with hardware requirements, obtained (usually by complex multi-licensing arrangements), installed, configured and tested. Finally, in order to test the system at anywhere near operating capacity, many human testers may need to be analyze and exercise the system during alpha and beta testing.
Often, the task of adequately testing such systems is the single largest expense in designing and developing software, especially if a computing system developer acquires more computing resources than is necessary. For example, a software developer might commit a substantial investment of funds capital computing equipment only to find later the hardware either does not meet its requirements or the quantity of hardware is underutilized in its design of software-based computing products or resources.
Reliability and compatibility also are important aspects that software developers and users rely upon when selecting new software and hardware. As the numbers of computing devices of today's computing environments increase, computing development and operation costs increase as well. Complexity of such environments also extends the time for a new computing product to reach the market. Today's computing environments are burdened by the conventional infrastructure underlying the computing environments. For example, the conventional infrastructure is associated with undue costs and efforts necessary to procure, install, provision and maintain the infrastructure. The burden associated with the infrastructure of today increases non-linearly with the complexity of the computing devices and interconnectivity of such devices that comprise the infrastructure. Furthermore, the complexity also increases the need for unnecessary evaluation and re-evaluation of new technologies from different vendors that are added to the conventional infrastructure.
Thus, there is a need for a computing environment system that provides users and developers an appropriate amount of reliable computing resources which can be dynamically allocated or de-allocated. A dynamic computing environment according to the present invention overcomes the aforementioned problems and provides advantages over the prior art.